This project has two major objectives. The first is to identify structural and functional properties of cognitive neuronal networks in cortex of association (prefrontal and posterior parietal) during working memory. The second objective is largely methodological: to substantiate the coupling between neural activity and hemodynamic changes in working memory. Both objectives will be pursued in the monkey by the combined use of four minimally invasive and behavior-compatible recording methods: near-infrared spectroscopy (NIRS), surface field-potential (FP) recording, local field-potential (LFP) recording, and unit-activity recording. NIRS signals and surface FPs will be recorded simultaneously with epidural probes. Unit activity and local field potentials (LFPs) will also be recorded simultaneously by means of transdural microelectrodes. Based on certain assumptions of cognitive network architecture and the spatial resolution of each method, the four methods will be used in combination to test three specific hypotheses of neural activation and hemodynamic change in the regions of interest during the performance of two working-memory tasks, spatial delayed response (DR) and non-spatial delayed matching to sample (DMS). The analysis will focus on the neural and hemodynamic activity during the retention of a sensory stimulus in working memory. NIRS, FP, and unit data will be correlated with each of the variables most relevant to the specific hypotheses to be tested: cortical location, stimulus or memorandum, task, time of trial, and level of correct performance. In the study of neural-hemodynamic coupling in cognitive function, special emphasis will be placed on the correlations between NIRS signals and electrical manifestations of cell discharge. These correlations are expected to provide crucial information on the neuronal basis of functional imaging signals, such as those obtained by BOLD fMRI, in human cognition.